AWS Certified CloudOps Engineer Associate SOA-C03 Practice Question

Cluster placement groups pack instances inside the same Availability Zone and place them on adjacent racks connected through a high-bandwidth, non-blocking network fabric. This delivers the lowest possible latency and the highest throughput between all 12 instances, which is essential for the simulation. A spread placement group deliberately separates instances onto different racks (and can span AZs), sacrificing network performance for fault isolation. A partition placement group distributes instances across partitions; traffic within a partition is fast, but communication between partitions can take additional network hops, so aggregate latency rises. Creating two separate cluster placement groups in different AZs would double the network distance between the two groups, preventing the low-latency, high-bandwidth communication required by the application.